JPH0726851B2 - Fiber-optic Sagnac interferometer for measuring rotational speed - Google Patents

Description

Detailed Description of the Invention

The present invention relates to a fiber optic Sagnac interferometer for measuring the speed of rotation, in which two light beams originating from a light source and produced by beam splitting illuminate a fiber coil in opposite directions. The interference pattern which is produced and then recombined, the output signal of which is scanned by a detector device whose output signal corresponds to the light intensity of the interference pattern, with the aid of a phase modulator located in the fiber coil Two light beams are combined from the demodulated variable signal components and modulated by a periodic square signal having a fundamental frequency f ₀ = 1 / 2t ₀ , t ₀ corresponding to the transit time of each of the light beams through the fiber coil And in relation to the cyclic variation in the modulation shift, the first signal component is combined with the frequency f ₀ during the two operating modes to be 2 Resulting in alternating opposite phase shifts of the two counter-rotating light beams, and in addition, the second signal component has a duration and two times whose rise corresponds to a transit time t ₀ or an integer multiple of t ₀ , respectively. A staircase ramp signal having an amplitude increment that compensates for a non-inverse incremental phase shift of the light beam, the amplified photodetector output signal being amplified by the first synchronous demodulator clocked at frequency f ₀ . To the analog-to-digital converter and, after digitization, is integrated in a digital integrator and via a digital-to-analog converter and a driver amplifier. A non-reverse phase shift of the light beam is applied to the phase modulator via the phase compensating the second signal component, where the scale factor A positive circuit is provided, which demodulates the amplified photodetector output signal for mode inversion and transmits the analog scale factor correction signal to the digital-to-analog converter via an integrating amplifier, where it is provided by a crystal oscillator. A clocked processor provides control and synchronization signals for the first synchronous demodulator, analog-to-digital converter and digital integrator.

In order to measure non-reverse phase shifts in fiber optic Sagnac interferometers, also known as ring interferometers, the printed publication DE-A 1-3, 144, 144.
162 to a phase modulator located in the region of the input of the fiber coil, while having a certain amplitude and frequency
Re-adjusting non-reverse phase shifts, for example with reference to rotational speed, by providing a sawtooth voltage that can be reversed by f ₀ = 1 / 2t ₀ and, on the other hand, also shifts the phase shifts. However, it is known that t ₀ indicates the transit time of each light beam through the fiber coil. Gradient compensation for or resetting sawtooth voltage (phase ramp) corresponds to the rotational speed, and thus directly proportional to Δφ ₀ / t _0, φ ₀ is the phase shift is not a reverse direction, for example caused by rotational movement Show.
In practice, however, it is hardly possible to use the so-called phase ramp resetting method described in the DE printing publication, because inertial measurement of rotational speed is particularly reversible phase shift, resetting signal amplitude or gradient and Accuracy is required with respect to scale factor proportionality, which cannot be guaranteed in analog circuit optics.

US Pat. No. 4,705,399 (=
European patent A1-No. 0,168,292) discloses a notable further development of a lamp resetting method for fiber optic ring interferometers, which was mentioned in principle in said DE printing publication. It operates entirely digitally for signal evaluation and conditioning, so that the required signal accuracy is unique to automatic measurement range reversal, especially with regard to reversible phase shift and lamp reset signal accuracy. Can be obtained by the correction of the scale factor. The functional principle of the digital phase ramp resetting method for a fiber optic rotational speed sensor according to the prior art is first explained with reference to FIG. 14, which is in a simplified representation in said printed publication US Corresponding to the prior art according to FIG. 14 of patent A-4,705,399 (also Fiber Optic Gyros, 1986, SPIE Volume 719, HCLefevre). , Integrated Optics, “A practical soluti for fiber optic gyroscopes.
on for the Fiber-Optic Gyroscope) ", reference is made to the above references for details.

A schematic representation of FIG. 14 of a Sagnac interferometer corresponding to said prior art and having a closed fiber loop is:
A light source L, for example a laser, is shown, whose parallel light beam is split into two light beams via a beam splitter ST1 and is applied to the interferometer fiber coil FS in opposite directions. The interferometer fiber coil FS preferably comprises an optical monomode fiber. Beam splitter ST
1 acts simultaneously as a mixer to recombine the two light beams after traversing the fiber coil FS.
The interference signal of the two superposed light beams passes through the second beam splitter ST2 and its output branch AUS to the photodetector PD, which scans the intensity of the interference pattern. The use of Δφ ₀ to indicate the phase difference between two counter-propagating light beams in the closed fiber coil FS keeps Δφ ₀ = 0 unless there is non-opposite interference. Reference should be made to the mathematical relationship between the phase difference Δφ ₀ and the rotation speed, the photodetector (referenced to the optical power density at the input of the PD and the measurement sensitivity. There is also the fiber coil FS). By introducing a constant reverse bias to be applied to the two counter-rotating light beams,
The two counter-rotating light beams (light waves) are moved to the operating point with the highest sensitivity of the interferometer by the phase modulator PM (2n + 1).
It is described how the sensitivity of the interferometer can be increased so that it is accurate in such a way that it is periodically shifted by an angle of π / 2 and n is an integer. For this purpose, the phase modulator PM therefore firstly
Excited by ₁ (t), which is, for example, ± π / 2,3
, 2π, ..., Which results in a periodic phase shift of 2t ₀ , where t ₀ represents the transit time of the light wave in the fiber coil FS.

In the case of the circuit according to FIG. 14, the non-reverse phase shift from the negative feedback to the phase modulator PM is compensated by the so-called phase ramp signal, as described in said US patent, and its Gradient is Δφ / t
It is directly proportional to ₀ , ie, Δφ, which is not a reverse phase shift. In the case of the circuit according to FIG. 14, the resetting phase ramp signal is a sawtooth or step signal, in which the height of the sawtooth amplification or the rising edge is equal to Δφ and the sawtooth or rising connection is The time corresponds to the running time t ₀ or an odd multiple of t ₀ .

Compared with analog solutions, this digital phase ramp principle described in the literature, in particular in the aforementioned US Pat. No. 4,705,399, is based on the principle that the scale factor of the phase modulation control signal and the phase ramp resetting signal is at least the principle. Has the decisive advantage that it is compensated in and that an exact definition of the sawtooth amplification of the reset signal is guaranteed.

The functioning of the circuit according to FIG. 14, namely to generate a stepped ramp resetting signal with alternating phase shifts and scale factor regularization alternating for frequency f ₀ and optimizing the operating point. Are briefly described below.

The output signal VD of the photodetector PD is raised via an impedance converter and an amplifier A ₀ and its output signal VD 'feeds a synchronous demodulator SYNCD which at a scanning frequency f ₀ = 1 / 2t ₀ . Be synchronized. The demodulated output signal passes through amplifier A as signal VA, which is generally combined by a filter into an analog-to-digital converter AD whose digital output signal is
It contains an item of code information that is directly proportional to the speed of rotation and the direction of rotation. The signal VAD then passes to the circuit GSC, which essentially comprises a digital integrator and, as a combined signal consisting of a digital phase ramp reset signal and a digital phase modulation signal, the combined phase control signal VSC. To supply. Thereafter, the combined digital signal VSC is converted into an analog control voltage VC in a digital-analog converter DA,
Further, it is transmitted to the phase modulator PM as a control voltage via the driver amplifier AP.

Circuit GSC for generating the signal VSC
Contains a first memory M ₁ for intermediate storage of the digital signal VAD, the intermediately stored output signal SM ₁ of which passes to the first input e ₁ of the adder ADD, which is of the second type. Cascaded to memory M ₂ . The output signal VSC intermediately stored in the second memory M ₂ is
It is fed back to the second input e ₂ of DD and added to the digital value of the rotational speed signal. Therefore, the adder ADD
The output signal SADD of 1 corresponds to the rotation angle.

The circuit described thus far for generating the phase modulation signal and the phase ramp reset signal is synchronized and controlled by the central processor CPU via the bidirectional bus BC, which is connected to the crystal stable oscillator OSC. Connected, it supplies the frequency f ₀ .

The number indicates a particular rotation angle, for example an arc of 4 seconds, the bit of which, apart from the number of bits per word, corresponds to the capacity of the adder ADD, said adder at the output SL. The overflow signal is supplied to the central processor CPU via the bus BC. Depending on the overflow signal of the adder and in agreement with the clock signal of the oscillator OSC, the central processor CPU generates various control and synchronization signals, which enable the correction of the scale factor. Depending on the overflow of the ADD, a switching of the modulation shift between "mode A" and "mode B" takes place, which is for example ± π /
The case where the modulation shift of 2 is in the method in which the mode A is held and the phase shift of ± 3 / 2π is held in the mode B is the case. As mentioned in the cited document, the detected amplitude difference between operating conditions and different modulation shifts is a measure of the scale factor error in case of phase ramp signal overflow. In order to correct this scale factor error, the known circuit according to FIG. 14 is equipped with a demodulator SFC, which has a mode A (for example ± π / 2 phase shift) and a mode B.
A difference in the amplitude of the f ₀ modulation signal at (for example, a phase shift of ± 3 / 2π) is detected, and this demodulated signal is used as an analog correction signal SIA via the integrating amplifier IA to obtain the digital-analog converter DA. Transmit to compensated analog input e _M.

Briefly described with reference to FIG. 14, the gradient of the phase ramp is a measure of the reset speed of rotation, and in the return phase at a ratio of 1: 3 to obtain a correction signal for the scale factor. The functional principle of the digital phase ramp resetting method for a fiber optic rotational speed sensor, with modulation shift reversal, is practical in practical operation of such a parasitic rotational speed measuring device for the following reasons: Leading to difficult.

The reversal from "mode A" to "mode B" depends directly on the ramp value of the reset sawtooth signal and thus on the rotation angle of the gyroscope. In this regard, the traversing of the ramp may correspond to an angular increment of the arc of, for example, about 4 seconds. However,
The reversal of the modulation shift results in a gyroscope bias reversal, via the amplitude of the modulation signal being tripled, and via the latter parasitic in the sensitive signal path of the photodetector signal VD. These unwanted parasitic effects are shown in FIG.
4 by the dotted line and the coupling count K in the circuit according to 4. However, since the reversal of this modulation shift signal depends on the rotation angle, it leads to the dead zone (lock-in) of the gyroscope. As explained below,
It is also possible to show that this effect leads to non-linearity of the scale factor outside the lock-in zone.

Printing Publication US Pat. No. 4,705,3
As shown in FIG. 13 of No. 99, given the occurrence of scale factor error, the photodetector signal VD will be different in the case of mode A modulation shift or mode B modulation shift. There is an average strength I. This difference in strength is integrated in the integrating amplifier IA and supplies the analog correction signal to the input e _M of the digital-analog converter DA. The frequency of change in intensity is equal to the frequency of change from mode A to mode B and is therefore equal to the ramp signal return frequency, for example, the reversal from mode A to mode B is a ramp overflow, ie the adder ADD. This is because it is generated by the signal SL of. The frequency of this change is directly proportional to the speed of rotation, that is, according to the example described in said US patent, a return of 2π corresponds to an angular increment of the arc of 5 seconds, and with a return frequency of 1 Hz this is 5 ° / h of rotation. Corresponds to speed.

At low rotational speeds, the frequency of this change can be arbitrarily low, and for this reason the integrating amplifier IA rarely contains strength difference information. At arbitrarily long intervals thereafter, where the modes do not change, the integrating amplifier IA does not have this information, and any small electrical zero error at its input is the amount of its output, the scale factor correction signal. Causes SIA to drift. For this reason, at low rotational speeds, the scale factor is subject to drift of the integrating amplifier IA. It is correct that the percentage error of the rotational speed caused by the drift of the scale factor, expressed in absolute value of the rotational speed, is small due to the low rotational speed. However, firstly, a sudden spike in rotational speed becomes problematic, because then the scale factor is still "wrong", and thus at least the scale factor control circuit once again reaches steady state conditions. To some extent, high absolute error in rotational speed also occurs.

Now, by means of the electromagnetic compatibility measure (EMC measure), ie in particular the shielding (shown in FIG. 14 by shielding the connection line from the driver amplifier AP to the phase modulator PM), the signal and the voltage. It appears to be a good idea to try to eliminate or at least reduce the electromagnetic parasitics (disturbances due to the coupling factor K, see FIG. 14) by installing filters in the feed line. However, there is a special EMC problem with the known interferometer design according to FIG. 14, in which the signal VSC or VC, VC 'contains the modulation frequency f ₀ = 1 / 2t ₀ , which is either at the oscillator OSC or at the processor C.
It is generated via PU. However, the photodetector signal VD contains rotational speed information with the same frequency and phase angle. This signal is detected in the synchronous demodulator SYNCD. A group of circuits that generate a modulation amount of frequency f ₀ and a circuit portion that conducts a signal of the same frequency that is sensitive to rotational speed are closely connected in space and are generally sent from a common power supply. Have to be done. This clearly presents the danger that electromagnetic energy of frequency f ₀ will enter the sensitive signal path (signal VD) parasitically. A stop filter for f _{0 in} the signal line is not possible because the desired signal information is exactly at this frequency. Therefore, unwanted parasitics are, for example, only amplifier A ₀ and synchronous demodulator SYNC.
It can be reduced to some extent by shielding D against the rest of the circuit and by filtering their power supplies.

A few examples for parasitic susceptibility will immediately reveal particular problems to those skilled in the art. This spectral component due to the frequency f ₀ in the signal VC or VC ′ is generally located in the slight V range. By contrast, in the rotation speed signal VD, a small n
The voltage in the V range depends on the light output, the sensitivity of the detector and the scale factor of the gyroscope, which is 1
Corresponds to a rotation speed of ° / h. In the case of such large amplitude differences, undesired parasitics despite all possible EMC measures, as symbolized by the coupling coefficient K between the signals VC, VC 'and VD in FIG. The path is unavoidable, due to the different parasitic amplitudes originally
And thus, due to the different error in the signal VD corresponding to the rotational speed, strictly speaking, it depends on the operating mode A or B, since, as hypothesized, the latter has a modulation amplitude three times higher than the former. And, therefore,
This is because it has a stronger parasitic effect. With respect to rotational speed, these different parasitic amplitudes lead to different gyroscope zero point errors, which are biased by B _a or
Shown as B _b , which depends on the instantaneous modulation state mode A or mode B.

The ramp gradient of the reset signal is thus adjusted via the closed control circuit in a manner that is directly proportional to the sum of the true input speed of rotation D _e and the respective bias B _a or B _b , which is the modulation mode. It depends on whether there is A or B.

In FIG. 1 of the accompanying drawings, D _e + B _a > 0,
D _e + B _b > 0 and B _a > B _b are retained, ie different
An example with B _a or B _b is shown. If D _e = const
If so, a different ramp gradient is created depending on whether there is a mode A or a mode B, because the ramp gradient is in direct proportion to D _e + B _a or D _e + B _This is because the adjustment is performed in a manner that is directly proportional to _b .

In the example shown in FIG. 1, the gradient during mode B is flatter than in mode A because B _b <B _a is assumed. Accordingly, larger than for independent bias (B _a = B _b), the shutdown time t _b in the mode B as compared to operating downtime t _a in the mode A from mode. However, since B _a and B _b input rotation speed not only D _e affect lamp gradient in both modes, Parusudeyu modes - duty ratio (T _a / T _b) also depends on the D _e. The scale factor non-linearity dealt with above results from this.

On the other hand, the lock-in effect mentioned above occurs in the rotational speed range in which the sums (D _e + B _a ) or (D _e + B _b ) have different signs for the two modes. , That is, (D _e + B _a )> 0 or (D _e
It is time to keep + B _b ) <0. This case, which is quite practical due to the low rotational speed, is displayed in FIG. The ramp of the reset signal (illustrated without modulation for clarity of display) starts in mode A, for example with a positive slope, because it is assumed (D _e
This is because + B _a )> 0 is applied. When the lamp reaches the (upper) overflow range (overflow signal SL in FIG. 14), a switch is made to mode B. However, the control then triggers a negative ramp gradient, because it now keeps (D _e + B _b ) <0, ie the ramp slope changes its sign and therefore leaves the overflow region, And mode A continues again. However, the positive ramp gradient once again belongs to this mode A until another reversal to mode B occurs. Therefore, the control is "captured" and held, ie the interferometer arrangement is locked in. How quickly the "Zyza clamp" displayed between modes A and B in FIG. 2 changes in the overflow limit is determined by the speed of control, which is generally very high. The lock-in condition is retained for the input rotational speed range, in which certain inequalities are fulfilled, ie the lock-in range has a width of | B _a −B _b |.

In the description given thus far, the particular ramp value is taken to be the instantaneous value of the "unmodulated" ramp φ (t). This ramp value is directly proportional to the rotation angle and corresponds to the signal SADD. The ramp gradient dφ / dt corresponds to the time derivative of the rotation angle, that is, the rotation speed. To clarify what should be the problem, the problem of the known lamp resetting method is that the lamp value, ie the value of the angle of rotation, has either operating mode A or B and thus a different bias B _a or B _b . , And that is because B _a ≠ B _b affects the ramp gradient itself. As displayed, the scale factor error information is the shape of the signal of change in intensity, whose frequency is directly proportional to the rotational speed,
And can vary between 0 and about 100 kHz, when the theoretical pulse-to-duty ratio of the modulation shift between mode A and mode B is 1: 3. However, the obvious advantage that changes in information-carrying strength occur more often at high rotational speeds and, as a result, better scale factor error information is at higher rotational speeds, is that the associated signal processor is relatively high. Against the fact that the frequency range has to be dealt with. However, as indicated, in the region of zero rotational speed, on the other hand, the known ramp resetting method leads to a lack of scale factor error information and thus to the problem of drift.

The object of the invention is therefore to avoid dead zones of the fiber optic rotation speed sensor in the case of low rotation speeds, ie the indicated lock-in effect, and to handle for the purpose of controlling the scale factor. Is to get a signal that is easy.

As a preliminary step towards achieving the target setting, as already explained with reference to FIGS. 1 and 2, the automatic change in the deviation of the modulation as a function of the angle of rotation causes the non-linearity of the scale factor. It has been recognized that and the range is a major cause for the lock-in effect ramp-over of the known method of wide frequency range of scale factor error information as a function of rotational speed. The present invention is therefore directed towards the goal of developing a modulation method for establishing a ramp resetting method, which, on the one hand, shifts the modulation from mode A to mode B by the order of obtaining scale factor error information. And, on the other hand, defining this inversion with a specific, eg fixed, timing and no longer as a function of the angle of rotation.

Starting from the known lamp resetting method, as specified in the preamble of claim 1, in the case of a fiber optic Sagnac interferometer for measuring the rotational speed according to the generic concept presented at the outset. According to the invention, a (second) digital adder is arranged between a digital integrator and a digital-analog converter, which comprises a second
The input of is supplied from the first pulse generator with a modulation offset signal which can be reversed by a predetermined timing between the two operating modes, which is clocked by the frequency f ₀ and (the second ) Second digital adder
To the input of the phase shift ± π due to the first operation mode A
It consists in transmitting a signal corresponding to / 4 and a signal corresponding to a phase shift of ± 3 / 4π for the second operating mode B.

The invention and advantageous details are explained in more detail below in exemplary embodiments with further reference to the drawings.

Arrangement groups, circuit groups and signals already known from the description of FIG. 14 are referred to by the same symbols in FIGS. 3 and 4. The explanation of the functional relationships already known to the reader will not be repeated. The first step is to describe the mode of operation of the arrangement according to FIG.

The photodetector PD observes the light intensity I of the output signal of the interferometer and supplies a corresponding signal VD, which is amplified in amplifier A _{0 to} signal VD '. The light intensity signal I leads to (1) a change in intensity ΔI _{0 with} frequency f ₀ ,
And with the aid of the signal DEMR1, a deviation of the system of rotational speed, which is detected in the (first) synchronized demodulator SYNCD1, which is supplied by the processor CPU and is of the same frequency as the output signal VSY1.
The frequency f ₀ is the interferometer operating frequency, which is f ₀ = 1 /
2t ₀ , which is t ₀ = nx1 / c, where n = index of refraction of the light guiding material of the fiber coil FS,
1 = optical path length in fiber coil FS and c = speed of light in vacuum.

(2) The frequency of the modulation shift inversion signal SMODS leads to a change in intensity ΔI ₁ (see FIG. 12) between modulation modes A and B, which is explained in more detail below and in the second. A system deviation of scale factor, ie in an interferometer, which is generated in a pulse generator GEN2 represented as a generator and which can be synchronized by a gyroscope operating frequency f ₀ via a signal SYNC2 supplied by the processor CPU. Deviation in the phase modulation deviation of the electro-optical phase modulator PM. This difference in intensity ΔI ₁ is due to the signal SMO
(Second) synchronous demodulator SYNCD2 with the help of DS
, And the result is the signal VSY2, which is a measure of the scale factor error and is already in FIG.
4 via an integrating amplifier IA, as is known from the circuit according to 4., to the analog setting input e _M of the digital-to-analog converter DA.

For improved clarity, the first rotational speed control circuit will be described below, followed by the scale factor control circuit. Including information about. (1 again) Rotation speed control circuit Signal VS corresponding to system deviation of rotation speed control circuit
Y1 is amplified in amplifier A and provides signal VA. Amplifier A simultaneously includes a filter, which
The printed publication U.S. Pat. No. 4,705,399, which determines the technical control characteristics of the overall control circuit and is dimensioned according to the design procedures known in the control art, has already been dealt with more than once in this respect. Reference is made to the issue. In practice, the output signal VA will have a component of the time integral of the input signal VSY1. The signal VA is digitized in the analog-to-digital modulator AD and the resulting digital signal VAD is stored in the memory M1 and these processes are controlled by the processor CPU via the bus BC. As is already known, the output signal SM1 is applied to the input e ₁ of the adder, here represented as the first adder ADD1, for better differentiation of the further adder. The output signal SADD1 of the first adder ADD1 is sent to the memory M2, whose output signal SM2 is the further input e ₂ of the first adder ADD1.
, Which is in full agreement with the known circuit according to FIG. The first adder ADD1 represents, in connection with the control assigned via the memory M2 and the bus BC, a ramp signal generator, which generates a ramp signal or step signal SM2, the rising height of which Corresponds to the value SM1 and its step width corresponds to the light transit time t ₀ .

By limiting the word length of the digital processing device consisting of the first adder ADD1 and the second memory M2, the signal SM2 of the signal SM2 is exceeded when the specific output value of the first adder ADD1 is exceeded. A ramp overflow is achieved, which is represented schematically in FIG.
In this respect, it is possible to advantageously develop the overflow characteristics of digital binary codes.

This signal, shown as the second signal component SM2, reaches the known circuit according to FIG. 14 and is sent to the optical phase modulator PM via an electrical module which will still be described, And a phase shift π _R , which is essentially limited to the 2π phase range by the overflow described above (see Figure 5).

According to the invention, the essentially rectangular modulated signal SMOD is signal SM2 in the second adder ADD2.
To the combined modulated signal VS
Produces C. The modulation signal SMOD is the modulation signal generator MO.
Generated at D and has a gyroscope frequency F ₀ . The modulation signal generator MOD is synchronized by the processor CPU or the oscillator OSC via the signal SYNC1. The amplitude of the signal SMOD can be varied by a factor of 3, which is in a manner which will be explained in even more detail below, and which is obtained from the already mentioned second pulse generator GEN2.
Via ODS (see FIG. 6). Signal component SMOD (Fig. 7
(Reference) is added to the signal SM2 in the second adder ADD2, converted into analog in the digital-analog converter DA, and then sent to the optical phase modulator PM via the driver amplifier AP. This signal component SMO
D introduces therein a phase shift φ _M , which is essentially ± π / 4 or ± 3 / 4π (see Figure 7). Therefore, the amplitude of modulation φ _M is limited by the signal SMODS. In the meantime, the amplitude is ±
The state of π / 4 is called “mode A”, and is ± 3 /
In the case of a larger deviation in the modulation amplitude of 4π, “mode B” is set.

As shown in FIG. 3, the inversion of the modulation amplitude is, for example, two signals, SMOD1 corresponding to mode A described and SMOD2 corresponding to mode B.
Is generated in the generator GEN1 with the described amplitude and via the changeover switch SEL which is alternately controlled by the signal SMODS, the second of the second adder ADD2.
Can occur as provided as signal SMOD at the input of

The digital processing word width of the second adder ADD2 is designed such that the output signal VSC always corresponds to the sum of the input signal SM2 or SMOD without causing the consequence of overflow. Combined signal V
SC (see FIG. 8) is sent to the digital-to-analog converter DA in the manner already described and the voltage signal V
It is converted to C and amplified to form signal VC ', the value of which is essentially proportional to the value of signal VSC. The voltage VC 'causes a phase shift Δφ _C in the already phase-shifted PM (see again FIG. 8).

Therefore, a circuit adapted to FIG. 3 and modified according to the invention is based on the following principle.

Unconverted digital ramp values are limited for the phase angle range 2π or ± π by truncating more than the significant bits π. Second adder ADD
A square signal is applied to the restricted ramp in 2,
It is the first pulse generator GEN1, SEL or M
Generated at OD and having a modulation frequency f ₀ ,
It is invertible in amplitude according to phase angle values of ± π / 4 (mode A) and ± 3 / 4π (mode B). This inversion is a second pulse generator GE whose frequency and pulse-duty ratio are freely selectable.
It is controlled by N2. It can be seen that the reversal value can be invariably predetermined by the processor CPU or can be controlled as a function of the rotation speed and tuned with the modulation frequency f ₀ .

The modulated ramp signal represented in FIG. 8, which is sent via the interferometer to the phase modulator PM, produces an intensity signal I (t) at the output of said interferometer, which is essentially The following relationships are satisfied.

I (t) to cos (Δφ (t) + Δφ
₀ (t)) There, Δφ (t) = Δφ _C (t) −Δφ _C (t−t ₀ ) is effective, and Δφ _C (t) is the phase shift in the phase modulator PM, and Δφ ₀ (T) is the Sagnac phase shift, which is proportional to the rotational speed and

[0040]

[Equation 1] Given by. Here, 1 is the fiber coil FS
Denotes the length of the optical path at, and D denotes its average diameter, λ is the wavelength of the light used, c is the speed of light in the open space and Ω is the speed of rotation.

The signal Δφ (t) is represented in FIG. Figure 1
0 is the resulting variation in intensity Δφ (t)
And as a function of Δφ ₀ (t). These two
As can be seen from the two figures, the phase shift Δφ ₀ (t) proportional to the rotational speed is such that the intensity I rises in the time intervals “I” and “IV” and the time intervals “II”, “I”.
Leads to the fact that it descends at "II" or "V".
The resulting variation in intensity with time I (t) is represented in FIG.

The difference in intensity ΔI ₀ is, as described above,
Detected with the help of the first synchronous demodulator SYNCD1 and serving to control the ramp gradient of the signal φ _R (see FIG. 5), the effect of Δφ ₀ is compensated by the average value of the phase shift Δφ, which is the same. , But with different signs. Then, for intensity I (t), keep I (t) -cos (Δφ (t)) = const. From this it follows that the ramp gradient of φ _R is a measure of rotational speed.

Instead of the rotation speed φ _R , it may be advantageous to read the rotation angle as its integral. This means that, for example, the ramp signal S at the output of the second memory M2
Output signal SADD of M2 or of the first adder ADD1
This can be done by counting the above-mentioned overflow of 1. Overflow and the resulting return at a phase angle φ _R of only 2π are in this process

[0044]

[Equation 2] Corresponds to the angular increment of. The diameter of the fiber coil is D = 6c
When an interferometer with m is used, the angular increment α
Is approximately 4 angular seconds. This overflow information can be provided to the processor CPU, for example by the first adder ADD1 via its overflow output signal SL and the bus BC, which counts the angular increments (summing them). (Again 2) Scale Factor Control Circuit As described above, the second synchronous demodulator SYNCD2
The output signal VSY2 of is a measure of the scale factor error. This error signal is fed to an integrating amplifier IA, which represents a filter that is dimensioned from a control engineering point of view and determines the control engineering characteristics of the scale factor control. The output signal SIA contains at least one component corresponding to the time integration of the input signal VSY2. Also, as in the known circuit according to FIG. 14, the signal SIA is fed to the input e _M of the digital-to-analog converter DA. The digital-to-analog converter DA may be represented as a multiplying digital-to-analog converter, ie its output quantity VC is essentially proportional to the product of the two signals VSC (digital) and SIA (analog). To do. Therefore, with the help of the information about the difference in the intensity ΔI ₁ (FIG. 12), the voltage shift of the signal VC, the modulation phase shift in the optical phase modulator PM is exactly ± π / 4 (mode A) or ± 3. / 4π (mode B). It is thus ensured that the return of the phase φ _R caused by the overflow (see FIG. 5) at the same time corresponds exactly to 2π.

FIG. 11 shows here operating points I, II, III, which determine the variation in intensity, given the presence of scale factor errors, for example in the form in which the optical phase shift appears to a lesser extent than desired. IV or V
How is the operating point, I ', II', III ', I
V ', V'shifted and intensity ΔI ₁ (Fig. 12)
Represents how the above changes in are produced. By sending the appropriate correction signal SIA to the input eM of the digital-to-analog converter DA, the output amplitude of the signal VC and thus the optical phase shift will be the desired operating point,
I, II, III, IV, V are reached again and increased so that the change in intensity ΔI ₁ disappears.

For completeness, FIG. 13 further shows the synchronization signal DEMR1 to be sent via the bus BC to the first synchronous demodulator SYNCD1.

The (partial) circuit represented in FIG. 4, in its mode of function, corresponds perfectly to that according to FIG. The only difference is that the second synchronous demodulator now has four inputs to its positive or negative input, in a manner similar to FIG. 14 of the printed publication US Pat. No. 4,705,399. The signal VS1, VS2 or VS4, VS3 of the sample-and-hold circuit is formed by two differential amplifiers A14 and A23, respectively, provided in a manner not represented in more detail. However, the signal of the change in intensity is finally fed to the integrating amplifier IA, via which the changeover switch K on the output side is connected.
2 is controlled by the freely selectable but fixed output signal SMODS of the second pulse generator GEN2,
That is, it differs from known circuits in that it is invariant to a predetermined inversion of the mode shift or inversion of the modulation shift.

It is true that in a circuit modified according to the invention, eg adapted to FIG. 3 or FIG. 4, the interference of different modulation shifts will continue to result in a change in bias. However, the latter is no longer a function of rotational speed, but rather its value is constant,
And corresponds to the average of the bias values of the two modulation shifts, weighted with respect to the duty ratio. Since the errors in frequency and pulse-duty ratio, by which the changes in the modulation shift are controlled and by which the scale factor error information occurs, are freely selectable within wide limits and are known among others: Scale factor control can be optimized for a fixed frequency.

In contrast to the known implementations for an optical factor inertial rotational speed sensor with digital phase ramp resetting, the present invention is based on the described lock-in effect and the scale factor associated therewith. Non-linearity is provided, providing the decisive advantage that the signal, which is more advantageous for the process, is for scale fiber control.

Detailed EMC measures previously retained as needed for gyroscope electronics can be avoided, at least in part.

The extra cost on the circuit needed in the digital part of the control electronics is relatively low and is not significant in the use of modern integrated circuits.

Since the modulation range of the phase modulator PM is slightly increased, this also means that there is a somewhat higher capacity for modulation in the analog driver component and only one bit resolution of the digital-to-analog converter DA. Need an increase in. However, given the considerable practical advantages of the implementation according to the invention, this slight extra cost is negligible.

If what has been stated above and in the claims above is the basic or interferometer operating frequency f ₀ = 1 /
Starting from 2t ₀ , and to that extent, it can be stated explicitly that the device according to the invention for measuring the rotational speed can also operate only with a slightly detuned frequency. it can.

Generally, within the framework of the invention, the central processor (CPU) can and is implemented by a commercially available microprocessor. Implementations are possible without departing from the inventive concept and also in "hardwired" logic.

[Brief description of drawings]

FIG. 1 shows the ramp of the resetting signal the already described change in the gradient, which change depends on the change in the deviation of the modulation, ie the mode of operation.

FIG. 2 shows the generation of the lock-in effect likewise already explained, in the case of low rotational speeds.

FIG. 3 shows a first exemplary embodiment of an interferometer arrangement according to the invention for measuring rotational speed.

FIG. 4 shows a modified embodiment of the arrangement according to FIG.

5 shows signal shapes to explain the mode of operation of the circuit according to FIGS. 3 and 4. FIG.

FIG. 6 shows signal shapes to explain the modes of operation of the circuit according to FIGS. 3 and 4.

FIG. 7 shows signal shapes to explain the modes of operation of the circuit according to FIGS. 3 and 4.

FIG. 8 shows signal shapes to explain the mode of operation of the circuit according to FIGS. 3 and 4.

FIG. 9 shows signal shapes to explain the modes of operation of the circuit according to FIGS. 3 and 4.

FIG. 10 shows signal shapes to explain the modes of operation of the circuit according to FIGS. 3 and 4.

FIG. 11 shows signal shapes to explain the modes of operation of the circuit according to FIGS. 3 and 4.

FIG. 12 shows signal shapes to explain the modes of operation of the circuit according to FIGS. 3 and 4.

FIG. 13 shows signal shapes to explain the modes of operation of the circuit according to FIGS. 3 and 4.

FIG. 14 shows a previously described arrangement, corresponding to the prior art, of a fiber optic interferometer with digital processing of the rotational speed signal and generation of a modulation shift signal and a ramp resetting signal.

[Explanation of symbols]

 AD analog-digital modulator ADD1 digital integrator ADD2 digital adder AM phase modulator AP driver amplifier CPU processor DA digital-analog converter FS fiber coil GEN1 first pulse generator GEN2 second pulse generator IA said part amplifier L light source M1 digital integrator M2 digital integrator PD detector device PM phase modulator SEL changeover switch SIA scale factor correction signal ST1 beam splitter SYNCD1 first synchronous demodulator SYNCD2 demodulator VA demodulated and amplified output signal VD output Signal VD 'Amplified photodetector output signal

Claims (6)

[Claims] 1. A light source (L) which emits light and splits a beam (S).
The two light beams generated by T1) are radiated in opposite directions into the fiber coil (FS) and are subsequently recombined, and the interference pattern produced is such that its output signal (VD) is the light of the interference pattern. Detector device (PD) corresponding to intensity
, The two light beams are combined from a plurality of variable signal components and have a fundamental frequency F ₀ = 1 / 2T ₀ with the aid of a phase modulator (PM) placed in the fiber coil. , Where T ₀ = modulated by a periodic square signal, where T ₀ = the transit time of each of the light beams through the fiber coil (FS), and in relation to the periodic variation in the deviation of the modulation, the first signal component is Between the two modes of operation (A and B) results in an alternating alternating phase shift (φ _M ) of the two counter-rotating light beams together with the frequency f ₀ , the second signal component The rise time of each is the running time t ₀
Or a stepped ramp signal having a splice time corresponding to an integer multiple of t ₀ and an amplitude increment that compensates for a non-incremental incremental phase shift of the two light beams, the amplified photodetector output signal (VD ′) ) Is a first synchronous demodulator (SYNCD) clocked by the frequency f _0.
1), the demodulated and amplified output signal (VA) is sent to an analog-to-digital converter (AD) and, after digitization, a digital integrator (M1, ADD1 and M).
2) and a phase modulator (PM) via a digital-to-analog converter (DA) and a driver amplifier (AP) for compensating the non-reverse phase shift of the light beam by the second signal component. ), And amplified by the photodetector output signal (V
A scale factor correction circuit is provided for demodulating D ') and transmitting an analog scale factor correction signal (SIA) to a digital-to-analog converter (DA) via an integrating amplifier (IA), and wherein a processor is provided. (CPU) is a first synchronous demodulator (SYNC)
D1), providing control and synchronization signals for the analog-to-digital converter (AD) and the digital integrator (M1, ADD1 and M2), where the digital integrator (M1, ADD1 and M2) and the digital- A digital adder (ADD2) is arranged between the analog converters (DA), at the second input between the first pulse generator (GEN1) and the two operating modes (A or B). Is provided with a modulation shift signal that can be inverted at a predetermined timing of the clock, which is clocked by the frequency f ₀ and is fed to the second input of the digital adder (ADD2) in the first operating mode ( A signal corresponding to the phase shift of ± π / 4 for A) and a signal corresponding to the phase shift of ± 3π / 4 for the second operation mode (B). Fiber optic Sagnac interferometer for transmitting and measuring rotational speed. 2. A pulse signal is located between the first input of the first pulse generator (GEN1) and the second input of the digital adder (ADD2) and the pulse signal passes to the second input of the digital adder (ADD2). Driven by the second pulse generator (GEN2) in accordance with the operation mode inversion in such a manner, the pulse signal has a phase shift of ± π / 4 in the first operation mode (A) and the second Operation mode (B)
The interferometer according to claim 1, further comprising a changeover switch (SEL) corresponding to a phase shift of ± 3π / 4 at. 3. The interferometer according to claim 2, wherein the operating frequency and the pulse-duty ratio of the second pulse generator (GEN2) are freely selectable. 4. The interferometer according to claim 2, wherein the pulse-duty ratio and / or the frequency of the second pulse generator (GEN2) can be controlled as a function of the rotational speed. 5. The pulse-duty ratio and / or frequency of the second pulse generator (GEN2) is the modulation frequency f _0.
3. The interferometer of claim 2, synchronized by. 6. The output signal of the second pulse generator (GEN2) is a demodulator (S) which supplies a scale factor correction signal.
Interferometer according to one of the preceding claims, which is the clock signal for YNCD2).